Should I take a statin? Your cholesterol might be high, you might have already had a heart attack. Your doctor might advise you to take a statin – cholesterol lowering medication. How do you make an informed choice?
Well, here’s the logic and outcomes used in the trials which evaluate if these little pills are indeed worth taking. It’s not my place to prescribe or make a decision for you. That decision always should come down to the individual after considering the possible benefits and possible harms. They will decide if it’s a good idea for them.
In this piece I’m looking at exactly how you can make that decision. If you take medications, or know significant others who do, then this could be an important bit of reading for you.
Here’s the seven bits you need to know before you decide to take a pill. This applies to any pill, not just statins.
1. How drug trials are run and why
The standard for understanding the effectiveness of a drug is the randomised controlled trial (RCT). In an RCT participants in the study (sometimes also called subjects) are randomly assigned to the treatment (getting the drug) or the placebo control group (get a sugar pill which isn’t the drug).
Both groups take their “medication” for some time and then we look at how things differed between the groups. If a lot less “events” (like heart attacks) happen in one group over another, then we might be prepared to judge the treatment as either beneficial or harmful depending on how things happen.
If, say, the treatment group has less heart attacks than the control group, that means the drug is somehow reducing the chance of a heart attack. If more people in the treatment group get, say, cancer then the treatment is harmful.
It is important to realise that you can get benefit and harm at the same time – e.g. reduced heart attack, but increased cancer.
If the drug is beneficial then the researcher, clinician, and you the consumer then have to make a judgement about whether it’s worth taking the drug for the benefit and possible harms. For example, how many people will benefit from taking the drug? How many people suffer adverse events (harm)?
2. What happens in a typical statin trial?
Here’s how it might work. You eventually end up with say 2000 people (with raised LDL-C cholesterol) in the trial. You randomise half to drug treatment (1000) and the other half are placebo control.
After 5 years you follow up the group and discover that 10 people in the treatment group (they got the statin) had a heart attack, 20 people in the control had a heart attack.
So that sounds good right? Yes, of course, half the number of people in the statin (treatment) group had heart attacks. So that would usually be expressed by the change in relative risk of having a heart attack. If we do that we can say that taking the statin drug reduces your chance of a heart attack by 50%. And that’s the way it is generally reported in the scientific literature and explained to you by your doctor.
That’s only one way to think about the beneficial effects. There are several more. Read on.
3, What about expressing the outcome as absolute risk?
Presenting people with the same numbers in a totally different way, called absolute risk, might make you think differently. What these same numbers mean is that if you are in the control group you have a 98% chance of NOT having a heart attack over five years, and if you take the stain you have a 99% chance of not having a heart attack in the next five years. In other words, the drugs gives you a 1% decrease in your chances. Sounds very different – would this affect your decision to take the drug?
4, Number needed to treat
Another way to express the same data is to think about how many people would have to take the drug for one person to benefit. In this case 1000 people took the drug for 10 less heart attacks. The number needed to treat (NNT) is 1000/10 = 100.
100 people need to take the drug for one person to benefit.
Does that do anything different to your decision making?
5. Adverse events?
So no matter which way you play with presenting the numbers, there are still 10 fewer heart attacks in the treatment group. These are 10 real people having 10 less heart attacks. Heart attacks cost money, aren’t fun and mean you will probably die earlier with more suffering than if you didn’t have one. So if there weren’t any harms through side effects then it’d probably still be a no brainer. We could give them to all sorts of people. We could consider adding it to the water supply (some people have suggested this!).
Some of the common harms assessed in statin trials are cancer, myopathy (muscle pain and poor function), and diabetes.
Using the same hypothetical trial above (we’ll look at actual trials in the table below later), let’s say 10 people in the statin group got cancer over the 5 years and 1 person in the control. 20 got diabetes in the statin group, 10 in the control, and 100 showed myopathy in the statin group and 30 in the control.
So the side effects in the statin group are generally higher for muscle pain, cancer and diabetes. This is a consistent finding in such trials.
Would these numbers change your mind about the benefits versus harms?
Again you could express all of these as relative risk (1000% increase in cancer, 100% increase in diabetes, 333% increase in myopathy). Or you could be more sensible and show a 0.9% increase in the chances of cancer, 1% in diabetes, 7% in myopathy.
In a third way, you could express this as number needed to harm. How many people need to take the statin to be harmed (suffer an adverse event)? The number needed to harm is 100 for cancer, 100 for diabetes, and 14.3 for the muscle pain. One in every 100 people taking the statin will get cancer, one will et diabetes and 1 in 14 will suffer muscle pain caused by the drug.
6. Putting it all together
The website thennt.com gives summary data for statins as well as other drugs for you to consider. It takes recent trials data and meta-analyses to figure this all out. This is a great place to start if you want a summary of what’s going on. I have copied their figures for statin benefit vs harm in both numbers and percentages for people who haven’t had a heart attack (called primary prevention) and second for people who have already had a heart attack (called secondary prevention).
Statin Drugs Given for 5 Years for Heart Disease Prevention (Without Known Heart Disease)
Benefits: Expressed as NNT
- None were helped (life saved)
- 1 in 104 were helped (preventing heart attack)
- 1 in 154 were helped (preventing stroke)
Benefits: Expressed as percentages
- 98% saw no benefit
- 0% were helped by being saved from death
- 0.96% were helped by preventing a heart attack
- 0.65% were helped by preventing a stroke
Harms: Expressed as NNT
- 1 in 50 were harmed (develop diabetes)
- 1 in 10 were harmed (muscle damage)
Harms: Expressed as percentages
- 2% were harmed by developing diabetes
- 10% were harmed by muscle damage
Statin Drugs Given for 5 Years for Heart Disease Prevention (With Known Heart Disease)
Benefits: Expressed as NNT
- 1 in 83 were helped (life saved)
- 1 in 39 were helped (preventing non-fatal heart attack)
- 1 in 125 were helped (preventing stroke)
Benefits: Expressed as percentages
- 96% saw no benefit
- 1.2% were helped by being saved from death
- 2.6% were helped by preventing a repeat heart attack
- 0.8% were helped by preventing a stroke
Harms: Expressed as NNT
- 1 in 50 were harmed (develop diabetes*)
- 1 in 10 were harmed (muscle damage)
Harms: Expressed as percentages
- 2% were harmed by developing diabetes
- 10% were harmed by muscle damage
7. Other issues: Sneaky reporting and sneaky experimental designs
Now that you know the basics of how these trials are run and how to interpret the results, here are a few things that are common in the way these trials are conducted which are worth considering in making your judgement about the results.
- Benefits are also always expressed as changes in relative risk (they sound good – e.g. taking the drug decreases your chance of a heart attack by 50%).
- Harms are almost always expressed as absolute risk, or the plain number (10 more developed cancer in the statin arm). This makes it hard to judge the benefit/harm against one another and minimises the discussion of the side effects.
- Statin trials often “run in periods” which effectively remove people who suffer adverse events in the first month. That’s correct, they want to take these people out of the equation so you can be excluded in the first month of the trial if you suffer immediate adverse events. The most likely of this is the muscle pain. So if anything, the trials underestimate the amount of adverse events. It is estimated that muscle pain affects about 40% of people who take statins, although there is no easy way to determine this under the “run in” conditions. Most of the large drug company funded trials don’t make their databases available.
- Adverse events are often explained away as an “anomaly”. I kid you not. For example “there is no known link between statins and cancer so the fact that the statin group had 10 times more cancer must just be accidental”.
- Adverse events are often grouped and analysed in ways which make it hard to tell what is going on. For example, muscle pain might only be considered harmful if it is 9 or more out of the 10 point scale. Groupings might be made so it is harder to detect statistical significance. For example, muscle pain might be put into 5 different “bins” and because numbers aren’t taken as a whole, the results appear “not significant” when in fact they are.
- There is evidence of adverse events in the following areas – muscle pain, cancer risk, cognitive functioning, and diabetes. Not all these are always measured, nor is the study necessarily set up to find these differences between groups for various statistical reasons. That doesn’t mean they will happen, but it is possible, and some trials have shown this to greater or lesser extents.
Look, I’m not the one who should prescribe or tell you whether you should take a statin or not. What I can help you do though is decipher what the benefits and risks are and help you ask your doctor the right questions about these.
They should be able to answer these questions and then you should be able to make a decision about the medications you should take.
These are ultimately your decisions, not your doctor’s.
May 12, 2015 · by ·
Should I take a statin? Your cholesterol might be high, you might have already had a heart attack. Your doctor might advise you to take a statin – cholesterol lowering medication. How do you make an informed choice?
Well, here’s the logic and outcomes used in the trials which evaluate if these little pills are indeed worth taking. It’s not my place to prescribe or make a decision for you. That decision always should come down to the individual after considering the possible benefits and possible harms. They will decide if it’s a good idea for them.
In this piece I’m looking at exactly how you can make that decision. If you take medications, or know significant others who do, then this could be an important bit of reading for you.
Here’s the seven bits you need to know before you decide to take a pill. This applies to any pill, not just statins.
1. How drug trials are run and why
The standard for understanding the effectiveness of a drug is the randomised controlled trial (RCT). In an RCT participants in the study (sometimes also called subjects) are randomly assigned to the treatment (getting the drug) or the placebo control group (get a sugar pill which isn’t the drug).
Both groups take their “medication” for some time and then we look at how things differed between the groups. If a lot less “events” (like heart attacks) happen in one group over another, then we might be prepared to judge the treatment as either beneficial or harmful depending on how things happen.
If, say, the treatment group has less heart attacks than the control group, that means the drug is somehow reducing the chance of a heart attack. If more people in the treatment group get, say, cancer then the treatment is harmful.
It is important to realise that you can get benefit and harm at the same time – e.g. reduced heart attack, but increased cancer.
If the drug is beneficial then the researcher, clinician, and you the consumer then have to make a judgement about whether it’s worth taking the drug for the benefit and possible harms. For example, how many people will benefit from taking the drug? How many people suffer adverse events (harm)?
2. What happens in a typical statin trial?
Here’s how it might work. You eventually end up with say 2000 people (with raised LDL-C cholesterol) in the trial. You randomise half to drug treatment (1000) and the other half are placebo control.
After 5 years you follow up the group and discover that 10 people in the treatment group (they got the statin) had a heart attack, 20 people in the control had a heart attack.
So that sounds good right? Yes, of course, half the number of people in the statin (treatment) group had heart attacks. So that would usually be expressed by the change in relative risk of having a heart attack. If we do that we can say that taking the statin drug reduces your chance of a heart attack by 50%. And that’s the way it is generally reported in the scientific literature and explained to you by your doctor.
That’s only one way to think about the beneficial effects. There are several more. Read on.
3, What about expressing the outcome as absolute risk?
Presenting people with the same numbers in a totally different way, called absolute risk, might make you think differently. What these same numbers mean is that if you are in the control group you have a 98% chance of NOT having a heart attack over five years, and if you take the stain you have a 99% chance of not having a heart attack in the next five years. In other words, the drugs gives you a 1% decrease in your chances. Sounds very different – would this affect your decision to take the drug?
4, Number needed to treat
Another way to express the same data is to think about how many people would have to take the drug for one person to benefit. In this case 1000 people took the drug for 10 less heart attacks. The number needed to treat (NNT) is 1000/10 = 100.
100 people need to take the drug for one person to benefit.
Does that do anything different to your decision making?
5. Adverse events?
So no matter which way you play with presenting the numbers, there are still 10 fewer heart attacks in the treatment group. These are 10 real people having 10 less heart attacks. Heart attacks cost money, aren’t fun and mean you will probably die earlier with more suffering than if you didn’t have one. So if there weren’t any harms through side effects then it’d probably still be a no brainer. We could give them to all sorts of people. We could consider adding it to the water supply (some people have suggested this!).
Some of the common harms assessed in statin trials are cancer, myopathy (muscle pain and poor function), and diabetes.
Using the same hypothetical trial above (we’ll look at actual trials in the table below later), let’s say 10 people in the statin group got cancer over the 5 years and 1 person in the control. 20 got diabetes in the statin group, 10 in the control, and 100 showed myopathy in the statin group and 30 in the control.
So the side effects in the statin group are generally higher for muscle pain, cancer and diabetes. This is a consistent finding in such trials.
Would these numbers change your mind about the benefits versus harms?
Again you could express all of these as relative risk (1000% increase in cancer, 100% increase in diabetes, 333% increase in myopathy). Or you could be more sensible and show a 0.9% increase in the chances of cancer, 1% in diabetes, 7% in myopathy.
In a third way, you could express this as number needed to harm. How many people need to take the statin to be harmed (suffer an adverse event)? The number needed to harm is 100 for cancer, 100 for diabetes, and 14.3 for the muscle pain. One in every 100 people taking the statin will get cancer, one will et diabetes and 1 in 14 will suffer muscle pain caused by the drug.
6. Putting it all together
The website thennt.com gives summary data for statins as well as other drugs for you to consider. It takes recent trials data and meta-analyses to figure this all out. This is a great place to start if you want a summary of what’s going on. I have copied their figures for statin benefit vs harm in both numbers and percentages for people who haven’t had a heart attack (called primary prevention) and second for people who have already had a heart attack (called secondary prevention).
Statin Drugs Given for 5 Years for Heart Disease Prevention (Without Known Heart Disease)
Benefits: Expressed as NNT
- None were helped (life saved)
- 1 in 104 were helped (preventing heart attack)
- 1 in 154 were helped (preventing stroke)
Benefits: Expressed as percentages
- 98% saw no benefit
- 0% were helped by being saved from death
- 0.96% were helped by preventing a heart attack
- 0.65% were helped by preventing a stroke
Harms: Expressed as NNT
- 1 in 50 were harmed (develop diabetes)
- 1 in 10 were harmed (muscle damage)
Harms: Expressed as percentages
- 2% were harmed by developing diabetes
- 10% were harmed by muscle damage
Statin Drugs Given for 5 Years for Heart Disease Prevention (With Known Heart Disease)
Benefits: Expressed as NNT
- 1 in 83 were helped (life saved)
- 1 in 39 were helped (preventing non-fatal heart attack)
- 1 in 125 were helped (preventing stroke)
Benefits: Expressed as percentages
- 96% saw no benefit
- 1.2% were helped by being saved from death
- 2.6% were helped by preventing a repeat heart attack
- 0.8% were helped by preventing a stroke
Harms: Expressed as NNT
- 1 in 50 were harmed (develop diabetes*)
- 1 in 10 were harmed (muscle damage)
Harms: Expressed as percentages
- 2% were harmed by developing diabetes
- 10% were harmed by muscle damage
7. Other issues: Sneaky reporting and sneaky experimental designs
Now that you know the basics of how these trials are run and how to interpret the results, here are a few things that are common in the way these trials are conducted which are worth considering in making your judgement about the results.
- Benefits are also always expressed as changes in relative risk (they sound good – e.g. taking the drug decreases your chance of a heart attack by 50%).
- Harms are almost always expressed as absolute risk, or the plain number (10 more developed cancer in the statin arm). This makes it hard to judge the benefit/harm against one another and minimises the discussion of the side effects.
- Statin trials often “run in periods” which effectively remove people who suffer adverse events in the first month. That’s correct, they want to take these people out of the equation so you can be excluded in the first month of the trial if you suffer immediate adverse events. The most likely of this is the muscle pain. So if anything, the trials underestimate the amount of adverse events. It is estimated that muscle pain affects about 40% of people who take statins, although there is no easy way to determine this under the “run in” conditions. Most of the large drug company funded trials don’t make their databases available.
- Adverse events are often explained away as an “anomaly”. I kid you not. For example “there is no known link between statins and cancer so the fact that the statin group had 10 times more cancer must just be accidental”.
- Adverse events are often grouped and analysed in ways which make it hard to tell what is going on. For example, muscle pain might only be considered harmful if it is 9 or more out of the 10 point scale. Groupings might be made so it is harder to detect statistical significance. For example, muscle pain might be put into 5 different “bins” and because numbers aren’t taken as a whole, the results appear “not significant” when in fact they are.
- There is evidence of adverse events in the following areas – muscle pain, cancer risk, cognitive functioning, and diabetes. Not all these are always measured, nor is the study necessarily set up to find these differences between groups for various statistical reasons. That doesn’t mean they will happen, but it is possible, and some trials have shown this to greater or lesser extents.
Look, I’m not the one who should prescribe or tell you whether you should take a statin or not. What I can help you do though is decipher what the benefits and risks are and help you ask your doctor the right questions about these.
They should be able to answer these questions and then you should be able to make a decision about the medications you should take.
These are ultimately your decisions, not your doctor’s.
May 12, 2015 · by ·
May 12, 2015 · by ·
No comments :
Post a Comment